The Daily Index — Friday, May 26, 2023 · TLDR Index

Across today’s newsletters

Today’s Index

TLDR

AI CANON

It is difficult for AI experts to keep up with everything new and even harder for beginners to know where to start. This post shares a curated list of resources for learning about modern AI. These papers, blog posts, courses, and guides have significantly impacted the field over the past several years.

a16z.com · 16 min

TLDR AI

OPENAI MAY LEAVE THE EU

OpenAI CEO Sam Altman said on Wednesday the company might consider leaving Europe if it could not comply with the upcoming AI regulations by the European Union.

archive.ph · 1 min

TLDR Web Dev

DEVELOPER’S GUIDE TO PASSWORDLESS AUTHENTICATION

Passwordless authentication has been rising in popularity. It is more secure and convenient than password-based authentication. This article gives a thorough deep dive of the different forms of passwordless authentication and shows you how you can implement it.

semaphoreci.com · 12 min

TLDR Crypto

TRYING TO QUALIFY FOR A GRAVITA AIRDROP

Gravita, a fork of the Liquity protocol, has introduced a new stablecoin, leading to speculation about a future airdrop. To maximize your potential airdrop eligibility, you can follow these steps: mint Gravita's Debt Tokens, contribute to the protocol's stability pool, and provide liquidity for GRAI on Uniswap. As a little background on the protocol, Gravita is seeking to overcome limitations of Liquity, accepting bLUSD and LSDs as collateral.

threadreaderapp.com · 4 min

TLDR Founders

NVIDIA WORTH 250X MORE IN 10 YEARS

With enthusiasm for machine learning, investor appetite for AI exposure, Nvidia stock surged nearly 30% after posting earnings - the single largest value change in a stock ever in dollar terms.

tomtunguz.com · 2 min

TLDR

THE STATE OF GPT

This Twitter thread contains notes from the Andrej Karpathy presentation on May 24 about the state of GPT, which discussed the GPT Assistant training pipeline, LLM psychology, and best practices.

threadreaderapp.com · 20 min

TLDR AI

MICROSOFT CALLS FOR AI REGULATION

Microsoft Corp. is calling for a new US agency to regulate artificial intelligence and licensing requirements to operate the most powerful AI tools, company President Brad Smith said Thursday.

archive.ph · 2 min

TLDR AI

ENHANCING DIFFUSION MODELS WITH REINFORCEMENT LEARNING

This paper investigates the use of reinforcement learning to optimize diffusion models for downstream objectives like image quality. The proposed approach, Denoising Diffusion Policy Optimization (DDPO), proves effective in adapting text-to-image models to challenging objectives and improving prompt-image alignment without additional data or human annotation.

arxiv.org · 18 min

TLDR AI

GOAT: SMALL MODELS BEAT GPT4 ON ARITHMETIC

To be fair, calculators also beat GPT4 on arithmetic. The interesting discovery here is that the llama tokenizer (which splits numbers into individual digits) provides benefits for arithmetic even over other similarly powerful base models fine-tuned in the same task.

arxiv.org · 23 min